| 摘要 |
Abnormality causes are automatically identified during daily quality control, based on the focused consideration of complex uncertainty factors and, especially, of the causes of device-side abnormalities (i.e., abnormalities of the optical system and the dispenser mechanism), the latter of which are often difficult to identify. The analyzer performance that affects measurement results can be estimated from analysis parameters and calibration results. Thus, uncertainty estimates are automatically calculated for each analysis item during quality control, and the estimates are compared with uncertainties obtained during actual QC sample measurement, thereby monitoring and evaluating the analyzer performance. Also, measurements are performed on QC samples of multiple concentrations that contain substances known to subject to particular influences such as those of the optical system, sample dispenser, and reagent dispenser, so that the causes of abnormalities can be identified. Uncertainty estimates calculated from the parameters set for the analysis items are compared with uncertainties obtained from the QC sample measurements. The data sets that have value deviations are classified on an item-by-item basis, and the classification pattern is used to determine which analyzer component is abnormal. |
